File size: 2,609 Bytes
a137e28
 
280d697
a137e28
de83e40
a137e28
de83e40
280d697
de83e40
280d697
a137e28
280d697
a137e28
 
 
280d697
 
a137e28
 
 
 
280d697
 
 
 
 
 
 
 
 
 
 
 
a137e28
 
 
 
 
44318e4
280d697
 
 
a137e28
280d697
a137e28
 
280d697
 
 
 
 
 
 
 
a137e28
 
 
 
280d697
 
a137e28
 
 
 
 
 
 
 
 
 
 
 
280d697
a137e28
 
ab260d7
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
---
license: mit
base_model: ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-2
tags:
- alignment-handbook
- dpo
- trl
- selm
datasets:
- HuggingFaceH4/ultrafeedback_binarized
model-index:
- name: SELM-Phi-3-mini-4k-instruct-iter-3
  results: []
---



<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->



Self-Exploring Language Models: Active Preference Elicitation for Online Alignment.



# SELM-Phi-3-mini-4k-instruct-iter-3



This model is a fine-tuned version of [ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-2](https://huggingface.co/ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-2) using synthetic data based on on the HuggingFaceH4/ultrafeedback_binarized dataset.



## Model description



- Model type: A 3.8B parameter Phi3-instruct-based Self-Exploring Language Models (SELM).
- License: MIT



## Results



|                                        | AlpacaEval 2.0 (LC WR) | MT-Bench (Average) |
|----------------------------------------|------------------------|--------------------|
| [SELM-Phi-3-mini-4k-instruct-iter-3](https://huggingface.co/ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-3)  |    &emsp; &emsp; &emsp;&emsp;           27.98          |   &emsp; &emsp; &emsp;         8.32       |
| [SELM-Phi-3-mini-4k-instruct-iter-2](https://huggingface.co/ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-2)  |    &emsp; &emsp; &emsp;&emsp;           26.79          |   &emsp; &emsp; &emsp;         8.44       |
| [SELM-Phi-3-mini-4k-instruct-iter-1](https://huggingface.co/ZhangShenao/SELM-Phi-3-mini-4k-instruct-iter-1)  |    &emsp; &emsp; &emsp;&emsp;           27.33          |   &emsp; &emsp; &emsp;         8.37       |
| [Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct)  |    &emsp; &emsp; &emsp;&emsp;         23.05         |  &emsp; &emsp; &emsp;         8.12       |


### Training hyperparameters

The following hyperparameters were used during training:
- alpha: 0.001
- beta: 0.01
- train_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- num_epochs: 1

### Framework versions

- Transformers 4.40.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.19.1

Paper: arxiv.org/abs/2405.19332